History

AGNOSTIC was not built in a day, but hours passing by were as many bricks which were laid.

AGNOSTIC was not developed in a day nor was it a six-month project. Rather it emerged from hands-on investment management, falling into traps and climbing out, independent thinking, sorting through the chaff, moments of serendipity and long maturation.

Encelade Ltd is a Swiss company founded in 2008. The Company’s activities result in the application of state of the art research in statistics. As a company formerly active in the field of Alternative Asset Management, Encelade has carried out investments through careful automatized and systematic implementation of short term quantitative trading strategies.

Early in the century it became clear that exponential growth in computing power would open wide the access to exploring a universe of investment strategies and subsequently to exploiting the best ones. Very quickly we understood that persistence in trading strategy performance was a key insight. Using momentum on strategies was an excellent way to exploit this persistence, though it called for refinement on how painful false discoveries could be addressed.

A number of methodologies were identified. Among others, White Reality Check¹ or Family Wise Error Rate. Although conceptually sound, these statistical tools suffer from an over-conservative bias, which more often than not leads to discard results as irrelevant. Separately, adding insult to injury, the persistence we had identified in the performance of trading strategies was deteriorating.

We understood that a new toolbox was needed. One able to:

• Profit from opportunities early, well before even the best statistical tools could identify them.
• Deal with a large set of possible choices, without falling into the trap of overfitting.
• Avoid the corollary binary 0/1 decisions resulting from frequentist statistical testing.
• Narrowly match the performance of the best available choice(s), i.e. the performance you usually get by backtesting and never obtain in the future.
• Dispense with assuming anything about the choices, i.e. be oblivious to the characteristics and detailed implementation of each particular choice.
• Consider realized performance as the only input.

And serendipity sets-off an unexpected chain of events. Parsing through old academic papers, our attention was caught by research on Universal Portfolios, published by Professor T. Cover at Stanford University back in 1991². Three key insights struck us:

• no assumption on the properties of the stocks in the portfolio,
• guarantees on performance with respect to the performance of the best stock,
• sequential decisions based only on past performance.

The paper does not directly deliver the new toolbox for a number of reasons. A major impediment is the reliance on an exponentially increasing number of paths that cannot be handled, despite the capacities of today’s computers. But the paper provided the inspiration on the direction for further research. An additional number of years reading papers mainly unrelated to finance³, combined with our own ideas and further research finally led to the birth of AGNOSTIC.

Initially intended for internal use, we quickly came to realize the algorithm’s capabilities could be safely shared and be of use to third parties for their own purposes; since AGNOSTIC does not need to know the intricacies of the choices offered to its wisdom, but only their past performance. This means AGNOSTIC can be used as a tool to come close to the best:

• trading strategy, among a selection of trading strategies;
• asset allocation model, among a set of acceptable ones;
• hedge funds to invest into, from a long-list of approved hedge funds;
• factors / risk premia / alternative beta, from a broader choice of implementable ones;
• parameters, among all possible combinations of parameters for a particular strategy;
• portfolio construction methodology, among all relevant ones, like mean variance, risk parity, etc.